Summary
PS depth imaging requires a model that flattens events on both PP and PS common image gathers (CIGs) and images corresponding events at consistent depths for both data types. To satisfy these constraints we build an objective function consisting of measures of gather flatness, depth consistency, and well information if available. Both PP and PS gathers are flattened simultaneously, and to ensure depth consistency a floating event constraint is implemented by penalizing the relative depth shifts between PP and PS images. The depth shifts are measured by Dynamic Warping (DW) because it suffers less from cycle-skipping than local calculations by solving the problem of matching two images globally and optimally. With well information, tomographic inversion is able to derive anisotropic velocity models from PP and PS reflection seismic data efficiently and accurately. Tests on synthetic data and a 3D land data example illustrate the effectiveness of the joint PP/PS tomography with DW.
Introduction
Multicomponent seismology has played an important role in the oil exploration industry in recent years. PS waves can provide valuable information in reservoir characterization, such as lithological discrimination or estimation of rock properties. Also, as shear waves are less affected by attenuation than compressional waves, they can help in imaging of reservoirs beneath gas clouds (Stewart et al., 2003). S-waves are usually more sensitive to anisotropy than P-waves and can provide an additional constraint on anisotropy estimation (Tsvankin, 2012). Tsvankin and Grechka (2011) and Cai and Tsvankin (2012) state that when both horizontal and dipping interfaces exist, combining PP and PS moveout can resolve both the vertical velocities and anisotropy. PS waves may also image the subsurface better where PS reflectivity is stronger than PP. With all the benefits of PS waves, it is crucial to build correct models to take full advantage of the combined information from PP and PS waves. To this end, joint tomographic inversion of PP and PS data is an excellent choice for its efficiency and flexibility for combining all constraints, as discussed in several publications (Stopin and Ehinger, 2001; Broto et al., 2003; Foss et al., 2005; Liu et al., 2006; Szydlik et al., 2007). Our algorithm for updating the model parameters, vertical P-wave velocity (VP0), vertical S-wave velocity (VS0), and Thomsen anisotropic parameters (e and d), is based on the gridded reflection tomography conducted on post-migration CIGs. In addition to correcting residual moveout (RMO) in image gathers, the algorithm measures the depth shifts automatically by DW and penalizes depth misties between the migrated PP and PS images. The inclusion of PS data and requirement for depth consistency introduce extra constraints and provide better effective illumination for model building.
